Article Preview
Top1. Introduction
Predicting resident intention in smart home based on the contextual modalities like activity, speech, physiological parameters, emotion, object affordances, etc., has become an integral part of Ubiquitous Computing. The smart home is an environment that is embedded with technologies based on Artificial Intelligence, Machine Learning, Deep Learning & Internet of Things. In the proposed research work the house appliances like door, television, fan, light, air conditioner, etc., are considered to illustrate the idea of resident intention prediction. These appliances are made smart components by embedding resident intention-based service recommending capability thereby increasing the satisfaction level of the resident. For example, based on the context: Resident is ‘Standing’ in front of door & issuing the voice command ‘OPEN’ the door has to understand that the intention of the resident is to open the door & it has to open automatically. The resident intention prediction system is context dependent & it involves thee important tasks to be performed: 1) Recognition of resident intention 2) Discover the suitable service 3) Recommend the service. Table 1 illustrates examples where the user intention is inferred based on contextual information.
Table 1. Resident intention prediction based on the context
| User | Context | Intention |
| Location | Activity |
| Resident | Living Room | Sitting & Manually on the TV | Turn on the television |
| Living Room | Sitting & Issuing Speech Command: ON | Turn on the television |
| Near the Door | Standing & Manually unlocking the Door | Open the main door |
| Near the Door | Standing & Issuing Speech Command: OPEN | Open the main door |
In this article, we survey novel research works on contextual modalities activity, speech, physiological parameters, object affordances and emotion along with corresponding machine learning algorithms like Recurrent Neural Network (RNN), Restricted Boltzmann Machine (RMM), Deep Neural Network (DNN), Convolutional Neural Network (CNN), Naive Bayes classifier, Latent-Structural Support Vector Machine (LS-SVM), Natural Language Processing (NLP), Deep Belief Networks (DBNs), Long Short Term Memory (LSTM), Support Vector Machine (SVM). The objective of this work is to survey appropriate findings that illustrates the use of technologies like Natural Language Processing (NLP), Artificial Intelligence (AI), Machine Learning (ML) and Internet of Things (IoT) for enabling resident-appliance interaction.